Developmental Considerations of the RDoC Initiative in Relation to Autism

• In Event: Early signals of risk for psychopathology: Developmental approaches informed by the Research Domain Criteria

Abstract

Introduction: The Research Domain Criteria (RDoC) initiative promised to revolutionize mental health characterization and treatment by acknowledging and addressing the conceptual and empirical limitations to our nosological system. The categorical approach, based in large part on clinical consensus, has failed to yield reliable biological mechanisms that instantiate and/or maintain pathophysiology operating in DSM defined disorders. Refining our characterization of phenotypic heterogeneity within and across DSM defined disorders remains a noble enterprise, but if we are willing to challenge the notion that DSM defined categories are natural kinds, or singular categories that exist in nature, then we must re-evaluate the biological plausibility of diagnostic thresholds. Novel sampling strategies, coupled with methodological and computational rigor, are required to address these challenges. Further a developmental reframing is also necessary to adjudicate concerns around the “reactive” nature of much mental health policy and practice.

Hypotheses: Novel sampling strategies, coupled with methodological and computational rigor, will elucidate early emerging phenotypic heterogeneity of high-risk profiles agnostic to DMS-defined categories.

Study Population: I will present data from two sampling strategies, one from a convenience community sample (n = ~1500), and one from family study in which we characterize brain and behavioral development in infant siblings of autistic children (n =~500).

Methods: Leveraging several examples from my lab, I will highlight our attempts to characterize 1) measurement invariance to ensure that we are measuring dimensions of behavior that have meaning, and psychometric integrity, across the typical-to-atypical continuum, and 2) data-driven derivation of risk profiles agnostic to DSM categories.

Results: In the first example, I show the potential of a data-driven approach, using multiple dimensional RDoC constructs, for population-based screening that elucidates multiple high-risk profiles of impairment, in contrast to the status quo binary/categorical screening approach. In the second example, I show that 1) leveraging longitudinal development and an unbiased ascertainment strategy, coupled with 2) a stringent test of measurement invariance, we identified an equal proportion of males and females showing early concerns for ASD-associated impairment. This latter finding contradicts the established 3/4:1 male:female sex ratio in autism. While the RDoC initiative took the first step in countering the universalist-natural-kind approach to psychiatric nosology, we believe that an even greater degree of epistemic humility (c.f. Hyman 2021) is warranted when considering the biological plausibility of traditional diagnostic thresholds. Further, our work addresses a common methodological failure that Paul Meehl referred to as the ubiquity of detached validity claims (i.e., the assumption that a construct validated in one context is equally valid in a new context). Novel sampling strategies coupled with methodological rigor promises to afford new opportunities to infuse the RDoC initiative with principles anchored in developmental science, perhaps moving the field toward a proactive precision medicine approach.

Author

• Jed T. Elison, University of Minnesota - Twin Cities
  Presenting Author